| The New Austrian Tunneling Method (NATM) is still the main excavation method for mountain tunnels currently. It takes advantage of classifications of rock mass and the monitoring data of construction to adjust the preliminary design so that it can form a dynamic information construction. The rational rock mass classification for tunnel design and construction plan is very important. Controlling the drilling and blasting parameters is not only related to the stability of tunnel surrounding rock and supporting structure, but also affects the construction schedule and cost. How to optimize the parameters of blasting process and guarantee the stability of the surrounding rock effectively are the key unsolved technical issues for drilling and blasting method at present.Large number of measured data should be collected inevitably in the course of monitoring construction of tunnel. It’s not easy to keep whole monitoring data not lost or confused with manual management. The disordered and deficient data will generate various abnormal curves which leads to difficulties and inconvenience when one analysis the changes of rock mass and supporting structure. Moreover, it also brings many difficulties for evaluating the time and space effects of tunnel in construction. Therefore, judging the stability of surrounding rock may lack scientific basis because of all drawbacks.In order to solve the problems above mentioned existed in the long distance tunnel, this paper will further study classification of surrounding rock, optimization of parameters of optical burst and information management of monitoring measurement, which combined with an under construction tunnel, Jixi to Yellow mountain in Anhui province. And based on field tests, laboratory tests, theoretical analysis and numerical simulation, an intelligent decision support system software will be produced in this paper, and it can easily guide a dynamic and effective design and construction and be adapted to the construction site.The paper has made the following innovative achievements:(1) Based on BQ method, the quickly classification of rock mass was proposed for satisfying the field site construction. The proposed method introduced the evolutionary theory of support vector regression establishs a nonlinear model between classification of rock mass and different classification indices.(2) In the presence of defects in the traditional method of least squares, the application of particle swarm optimization (PSO) to optimize exponential, hyperbolic and logarithmic functions of regression coefficients achieved an automatic fitting for any monitoring data. And also, the use of coupled nonlinear algorithm of PSO and BP neural network for establishing intelligent model of surrounding rock deformation and time accomplish advanced prediction of rock deformation; Besides, through using monitoring data of rock pressure as continuous interpolation of key points, the finite element method was introduced to calculate moment, shear force and axial force of the preliminary supporting structure by virtue of Load-Structure theory.(3) The constrained multi-objective programming model for optimizing parameters of smooth blasting is proposed. The object of parameters optimization is to minimize range of loosen and sizes of underbreak and overbreak, and to present a corresponding algorithm depended on the modern optimization theory on the premise that ensure the stability of surrounding rock after blasting.(4) The coupled algorithm of PSO and BP neural network for solving the constrained multi-objective programming of above mentioned problems is introduced with field tests and finite element simulation.(5) With visual programming techniques, the rapid classification and burst parameters optimization module of surrounding rock are developed and integrated into a system. And, we developed a high degree of intelligence and automation integrated software, "highway tunnel construction intelligent decision support system". And it has been successfully applied in the Buddha tunnel and JiNing highway Tunnel groups. |
PDF Full Text Request